CN110801203B - Human cranial nerve fiber tracking method based on local features - Google Patents

Human cranial nerve fiber tracking method based on local features Download PDF

Info

Publication number
CN110801203B
CN110801203B CN201911043684.2A CN201911043684A CN110801203B CN 110801203 B CN110801203 B CN 110801203B CN 201911043684 A CN201911043684 A CN 201911043684A CN 110801203 B CN110801203 B CN 110801203B
Authority
CN
China
Prior art keywords
interpolation
tensor
grid
human brain
points
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201911043684.2A
Other languages
Chinese (zh)
Other versions
CN110801203A (en
Inventor
毕重科
朱林
李欣昕
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tianjin University
Original Assignee
Tianjin University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tianjin University filed Critical Tianjin University
Priority to CN201911043684.2A priority Critical patent/CN110801203B/en
Publication of CN110801203A publication Critical patent/CN110801203A/en
Application granted granted Critical
Publication of CN110801203B publication Critical patent/CN110801203B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/40—Detecting, measuring or recording for evaluating the nervous system
    • A61B5/4058—Detecting, measuring or recording for evaluating the nervous system for evaluating the central nervous system
    • A61B5/4064—Evaluating the brain

Abstract

The invention discloses a human brain nerve fiber tracking method based on local features, which is used for solving the problem that the periphery of a singular point cannot be correctly tracked in the human brain nerve fiber tracking process. The method mainly comprises the following steps: the method comprises the following steps of firstly, finding all grids with singular points from a human brain tensor field by using a singular point positioning method; step two, judging whether the midpoint of the grid falls on four quadrant angle bisectors or the center, if so, removing the side corresponding to the tensor with smaller anisotropy during the interpolation of the feature vector, and otherwise, removing the side with the farthest distance; thirdly, tensor field interpolation is carried out, the characteristic value of the midpoint of the grid is calculated by utilizing a scalar bilinear interpolation method, and the characteristic vector of the midpoint of the grid is calculated by utilizing tensor interpolation; and step four, rendering an interpolation result. The human brain nerve fiber tracking method can accurately find singular points in the human brain tensor field and overcome the problem of inaccurate tracking around the singular points, thereby improving the tracking accuracy.

Description

Human cranial nerve fiber tracking method based on local features
Technical Field
The invention relates to the technical field of human cranial nerve fiber tracking, in particular to a method for solving the problem that the periphery of a singular point cannot be correctly tracked in the human cranial nerve fiber tracking process.
Background
There are many nerve fibers in the human brain, and these intricate neurons are interwoven together to form a network. The network of neurons controls different levels of consciousness, movement, sleep, etc., for example, the human brain coordinates our perception, thinking and movement through its neuronal activity. Therefore, the tracking of the human brain nerve fibers can help human brain researchers draw detailed human brain nerve circuit diagrams, and further promotes the cognition of neuroscientists on the human brain. However, the human brain nerve fiber tracking is not ideal because of the large quantity, complexity and randomness of movement of the human brain nerve fiber.
At present, the tracking of human cranial nerve fibers mainly adopts methods such as electric probes, chemical reagents, genetic modification, tensor field interpolation and the like. Mathematically, the human brain can be represented as a diffusion tensor field, and the neurons are tensors in the tensor field, so that the tracking of the human brain nerve fibers is the tracking of flow lines in the tensor field. In general, a motion trajectory can be accurately tracked by a tensor field interpolation method, but when there is a tensor point (singular point) where the anisotropy is zero, the brownian motion has almost no direction around the singular point. It is difficult for the prior art to accurately track around singular points.
Therefore, there is a need for a method for neuroscientists to accurately find singular points in the human brain tensor field and to overcome the problem of inaccurate tracking around the singular points.
Disclosure of Invention
Aiming at the prior art, the invention provides a human cranial nerve fiber tracking method based on local characteristics, which solves the problem that the periphery of a singular point cannot be accurately tracked in the prior human cranial nerve fiber tracking technology.
In order to solve the technical problem, the invention provides a human cranial nerve fiber tracking method based on local features, which comprises the following steps:
the method comprises the following steps of firstly, finding all grids with singular points from a human brain tensor field by using a singular point positioning method;
step two, judging whether singular points in the grid are located on four quadrant angle bisectors or centers, if so, removing the edges corresponding to the smaller tensor of the anisotropic FA during the interpolation of the eigenvector, and otherwise, removing the edges with the farthest distance;
thirdly, tensor field interpolation, namely calculating an eigenvalue lambda of an interpolation point in the grid by using a scalar bilinear interpolation method, and calculating an eigenvector e of the interpolation point in the grid by using tensor interpolation;
and step four, rendering an interpolation result.
Further, in the first step of the method of the present invention, the singular point positioning method comprises the following steps:
firstly, determining the principal directions of four vertexes of a grid, wherein an eigenvector corresponding to the maximum eigenvalue is the principal direction of the tensor; then judging whether cos alpha of the tensor rotation angle of any two adjacent vertexes is smaller than zero or not; and finally, calculating the number n of cos alpha smaller than zero, wherein if n is an odd number, singular points exist in the grid, and otherwise, the singular points do not exist.
In step two of the method of the present invention, the expression of the anisotropy FA is as follows:
Figure BDA0002253544300000021
wherein the content of the first and second substances,
Figure BDA0002253544300000022
λ1，λ2，λ3is the eigenvalue of the matrix.
In the third step of the method of the present invention, the eigenvalue λ of the interpolation point in the grid is expressed as follows:
Figure BDA0002253544300000023
wherein the content of the first and second substances,
Figure BDA0002253544300000024
and
Figure BDA0002253544300000025
is R1And R2Eigenvalues of the two tensor points.
In the third step of the method of the invention, the feature vector e of the interpolation point in the grid is expressed as follows:
Figure BDA0002253544300000026
wherein the coefficient SyIs composed of
Figure BDA0002253544300000027
Figure BDA0002253544300000028
Are each R1And R2The feature vectors of the two points are,
Figure BDA0002253544300000029
is a rotation matrix.
Compared with the prior art, the invention has the beneficial effects that:
first, the present invention provides a singular point locating method in step one, by which all grids with singular points can be accurately found in the human brain tensor field.
Secondly, the invention provides a method for eliminating the interference of singular points in the tensor field to the tracking in the step two, and the problem that the singular points can not be accurately tracked around is solved.
Thirdly, during the interpolation of the tensor field in the third step, the eigenvalue and the eigenvector are respectively interpolated, and as a result, the anisotropy of the tensor field is kept as much as possible.
In conclusion, the invention provides a method for solving the problem that the periphery of a singular point cannot be correctly tracked in the process of tracking the human cranial nerve fibers. The method improves the accuracy of tracking the human cranial nerve fibers and further promotes the cognition of human cranial nerve circuits by human cranial researchers.
Drawings
FIG. 1 is a flow chart of the human cranial nerve fiber tracking method based on local features according to the present invention.
FIG. 2 is a schematic diagram of locating singular points in the present invention.
Figure 3 is a schematic diagram of tensor field interpolation in the present invention.
FIG. 4 is a partial code diagram of eigenvalue interpolation in the present invention.
FIG. 5 is a partial code diagram of feature vector interpolation in the present invention.
FIG. 6 is a diagram of the present invention in which a point falls on a first bisector of the quadrant angle, and the left part of the code is removed by comparing the anisotropy magnitudes of the corresponding two vertex tensors.
FIG. 7 is a graph showing the tracking effect of the present invention in the vicinity of the triple singular point.
FIG. 8 is a graph showing the tracking effect of the present invention in the vicinity of the wedge singularity.
Detailed Description
The invention will be further described with reference to the following figures and specific examples, which are not intended to limit the invention in any way.
The invention provides a human brain nerve fiber tracking method based on local features, which mainly comprises the following steps: finding all grids with singular points from a human brain tensor field by using a singular point positioning method; judging whether singular points in the grid fall on four quadrant angle bisectors or centers, if so, removing the side corresponding to the smaller tensor of the anisotropy FA during the interpolation of the eigenvector, and otherwise, removing the side with the farthest distance; tensor field interpolation, namely calculating an eigenvalue lambda of an interpolation point in a grid by utilizing a scalar bilinear interpolation method, and calculating an eigenvector e of the interpolation point in the grid by utilizing tensor interpolation; and rendering an interpolation result, accurately tracking the nerve fibers near the singular point in the human brain tensor field, and simultaneously distributing the nerve fibers around the singular point in a uniform and accurate distribution mode.
As shown in fig. 1, the specific process is as follows:
step 101, finding all grids with singular points from a human brain volume field by using a singular point positioning method;
in data representation, a singular point is a tensor with zero anisotropy, which has two or more equal eigenvalues. For one tensor, the eigenvector e generally represents the direction of the tensor, and its magnitude is represented by the eigenvalue λ, so the eigenvector corresponding to the largest eigenvalue is the principal direction of the tensor. Since brownian motion has almost no direction of motion around a singular point, it is difficult to accurately track around the singular point.
The singular point positioning method comprises the following steps: firstly, determining the principal directions of four vertexes of a grid, wherein an eigenvector corresponding to the maximum eigenvalue is the principal direction of the tensor; then, whether cos alpha of the tensor rotation angle of any two adjacent vertexes is smaller than zero is judged. And finally, calculating the number n of cos alpha smaller than zero, wherein if n is an odd number, singular points exist in the grid, and otherwise, the singular points do not exist.
FIG. 2 is a schematic diagram of locating singular points in the present invention. The four vertices of each rectangle in the figure are the four tensor samples, with the arrows pointing in the principal direction of the tensor. The principal direction rotation angle cos α of the two vertex tensors corresponding to the thick edges in the graph is smaller than zero. If a is a thick edge and is an odd number, singular points exist in the grid; and b, if two thick edges are even numbers, singular points do not exist in the grid.
Step 102, judging whether singular points in the grid are located on four quadrant angle bisectors or centers, if so, executing step 103, and removing edges corresponding to the smaller anisotropy FA tensor during feature vector interpolation; otherwise, step 104 is performed to remove the edge farthest away.
Anisotropy is an important indicator of the study of the tensor field. At present, various anisotropy calculation methods exist, such as anisotropy index, fractional anisotropy, relative anisotropy and volume fraction, etc., however, the present invention calculates tensor anisotropy using the fractional anisotropy calculation method. The calculation formula of the anisotropy FA, i.e.
Figure BDA0002253544300000041
Wherein the content of the first and second substances,
Figure BDA0002253544300000042
λ1，λ2，λ3is the eigenvalue of the matrix.
There is a subtle difference in the comparison of anisotropy when the tensor points in the grid fall on the four quadrant bisector and center. When the tensor point falls on the four-quadrant angular bisector, comparing the anisotropy of the tensors corresponding to the two vertexes; when the tensor point falls on the center, the sum of the two tensor anisotropies for each edge should be calculated and then the magnitudes are compared. FIG. 6 is a partial code diagram of the invention with dots falling on the first quadrant bisector and the left removed by comparing the magnitude of anisotropy.
And 105, carrying out tensor field interpolation, calculating an eigenvalue lambda of an interpolation point in the grid by using a scalar bilinear interpolation method, and calculating an eigenvector e of the interpolation point in the grid by using tensor interpolation.
In order to preserve anisotropy of the tensor field as much as possible, the tensor field interpolation separates eigenvalues and eigenvectors, and calculates by a scalar bilinear interpolation method as shown in fig. 4 and a tensor interpolation method as shown in fig. 5, respectively. The eigenvalues λ of the interpolation points, R as shown in FIG. 31And R2Are two dots, R1Characteristic value of point
Figure BDA0002253544300000043
Namely, it is
Figure BDA0002253544300000044
Then R2Characteristic value of point
Figure BDA0002253544300000045
Namely, it is
Figure BDA0002253544300000046
The characteristic value λ of the point P, i.e.
Figure BDA0002253544300000047
The feature vector of the interpolation point is e. As shown in FIG. 3, first R1Characteristics of the dots(Vector)
Figure BDA0002253544300000048
Namely, it is
Figure BDA0002253544300000051
Then R2Feature vector of points
Figure BDA0002253544300000052
Namely, it is
Figure BDA0002253544300000053
Then the feature vector e of point P, i.e.
Figure BDA0002253544300000054
Wherein
Figure BDA0002253544300000055
For a rotation matrix, coefficient SxAnd SyAre respectively as
Figure BDA0002253544300000056
Figure BDA0002253544300000057
And step 106, rendering an interpolation result.
Two types of singular points mainly exist in a human brain tensor field, which are a trisection (Trisector) and a wedge (wedge), respectively, and the streamline is difficult to accurately track near the two types of singular points by using a plurality of existing tracking technologies, and the tracking results always have the problems of infinite closeness, intersection and the like. To verify the effectiveness of our proposed local feature-based human cranial nerve fiber tracking method, we chose to track around these two singular points separately. Fig. 7 and 8 are graphs of the tracking effect in the vicinity of the triple point and the wedge singularity, respectively. The points circled by the boxes in the figure are seed points, i.e. the starting points of the tracing of each streamline. Obviously, the streamline traced by each seed point is uniformly distributed near the singular point, and the problems of infinite proximity, intersection and the like do not occur.
While the present invention has been described with reference to the accompanying drawings, the present invention is not limited to the above-described embodiments, which are illustrative only and not restrictive, and various modifications which do not depart from the spirit of the present invention and which are intended to be covered by the claims of the present invention may be made by those skilled in the art.

Claims (4)

1. A human brain nerve fiber tracking method based on local features is characterized by comprising the following steps:
the method comprises the following steps of firstly, finding all grids with singular points from a human brain tensor field by using a singular point positioning method; the method comprises the following specific steps: firstly, determining the principal directions of four vertexes of a grid, wherein an eigenvector corresponding to the maximum eigenvalue is the principal direction of the tensor; then judging whether cos alpha of the tensor rotation angle of any two adjacent vertexes is smaller than zero or not; finally, calculating the number n of cos alpha smaller than zero, wherein if n is an odd number, singular points exist in the grid, otherwise, the singular points do not exist;
step two, judging whether singular points in the grid are located on four quadrant angle bisectors or centers, if so, removing the edges corresponding to the smaller tensor of the anisotropic FA during the interpolation of the eigenvector, and otherwise, removing the edges with the farthest distance;
thirdly, tensor field interpolation, namely calculating an eigenvalue lambda of an interpolation point in the grid by using a scalar bilinear interpolation method, and calculating an eigenvector e of the interpolation point in the grid by using tensor interpolation;
and step four, rendering an interpolation result.
2. The method for tracking human cranial nerve fibers according to claim 1, wherein in the second step, the expression of anisotropy (FA) is as follows:
Figure FDA0003372939520000011
wherein the content of the first and second substances,
Figure FDA0003372939520000012
λ1，λ2，λ3is the eigenvalue of the matrix.
3. The method for tracking human brain nerve fibers based on local features according to claim 1, wherein in the third step, the feature value λ of the interpolation points in the grid is expressed as follows:
Figure FDA0003372939520000013
wherein the content of the first and second substances,
Figure FDA0003372939520000014
and
Figure FDA0003372939520000015
are each R1And R2Eigenvalues of the two tensor points.
4. The method for tracking human brain nerve fibers based on local features according to claim 1, wherein in the third step, the feature vector e of the interpolation points in the grid is represented as follows:
Figure FDA0003372939520000016
wherein the coefficient SyIs composed of
Figure FDA0003372939520000017
Figure FDA0003372939520000018
Are each R1And R2The feature vectors of the two points are,
Figure FDA0003372939520000019
is a rotation matrix.
CN201911043684.2A 2019-10-30 2019-10-30 Human cranial nerve fiber tracking method based on local features Active CN110801203B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911043684.2A CN110801203B (en) 2019-10-30 2019-10-30 Human cranial nerve fiber tracking method based on local features

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911043684.2A CN110801203B (en) 2019-10-30 2019-10-30 Human cranial nerve fiber tracking method based on local features

Publications (2)

Publication Number Publication Date
CN110801203A CN110801203A (en) 2020-02-18
CN110801203B true CN110801203B (en) 2022-02-15

Family

ID=69489752

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911043684.2A Active CN110801203B (en) 2019-10-30 2019-10-30 Human cranial nerve fiber tracking method based on local features

Country Status (1)

Country Link
CN (1) CN110801203B (en)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002017707A (en) * 2000-07-05 2002-01-22 Ge Yokogawa Medical Systems Ltd Image pickup face determining method and mri device
WO2006106242A1 (en) * 2005-04-07 2006-10-12 Inria Institut National De Recherche En Informatique Et En Automatique Improved device for processing raw images or tensor images
CN101833790A (en) * 2010-04-30 2010-09-15 浙江大学 Method for generating anisotropic quadrilateral grid based on wave equations
CN103049901A (en) * 2012-08-03 2013-04-17 上海理工大学 Magnetic resonance diffusion tensor imaging fiber bundle tracking device
CN103700146A (en) * 2013-12-01 2014-04-02 北京航空航天大学 Three-dimensional data visualization enhancing method based on anisotropic structure tensor
CN103914431A (en) * 2014-01-16 2014-07-09 同济大学 Mesh-less method for calculating anisotropic structure radar cross section
CN105842642A (en) * 2016-03-17 2016-08-10 天津大学 Fractional anisotropy microstructure characteristic extraction method based on kurtosis tensor and apparatus thereof
CN110021003A (en) * 2019-02-14 2019-07-16 清华大学 Image processing method, image processing apparatus and magnetic resonance imaging device

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102005028475B4 (en) * 2005-06-20 2008-04-03 Siemens Ag Method and device for determining coefficients of a magnetic resonance diffusion tensor
US7480400B2 (en) * 2006-03-16 2009-01-20 Siemens Medical Solutions Usa, Inc. Detection of fiber pathways

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002017707A (en) * 2000-07-05 2002-01-22 Ge Yokogawa Medical Systems Ltd Image pickup face determining method and mri device
WO2006106242A1 (en) * 2005-04-07 2006-10-12 Inria Institut National De Recherche En Informatique Et En Automatique Improved device for processing raw images or tensor images
CN101833790A (en) * 2010-04-30 2010-09-15 浙江大学 Method for generating anisotropic quadrilateral grid based on wave equations
CN103049901A (en) * 2012-08-03 2013-04-17 上海理工大学 Magnetic resonance diffusion tensor imaging fiber bundle tracking device
CN103700146A (en) * 2013-12-01 2014-04-02 北京航空航天大学 Three-dimensional data visualization enhancing method based on anisotropic structure tensor
CN103914431A (en) * 2014-01-16 2014-07-09 同济大学 Mesh-less method for calculating anisotropic structure radar cross section
CN105842642A (en) * 2016-03-17 2016-08-10 天津大学 Fractional anisotropy microstructure characteristic extraction method based on kurtosis tensor and apparatus thereof
CN110021003A (en) * 2019-02-14 2019-07-16 清华大学 Image processing method, image processing apparatus and magnetic resonance imaging device

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
A survey on visualization of tensor field;Bi C, 等;《Journal of Visualization》;20190322;全文 *
Collateral nerve fibers in human spinal cord: visualization with magnetic resonance diffusion tensor imaging;Mamata H, 等;《Neuroimage》;20060120;全文 *
Interactive control of mesh topology in quadrilateral mesh generation based on 2D tensor fields;Bi C,等;《International Symposium on Visual Computing. Springer, Berlin, Heidelberg》;20121231;全文 *
Voxel based versus region of interest analysis in diffusion tensor imaging of neurodevelopment;Snook L,等;《Neuroimage》;20070101;第34卷(第1期);全文 *
基于k-medoids聚类的人脑DTI图像分割算法及其纤维追踪的研究;林涛;《中国优秀硕士学位论文全文数据库》;20170215(第2期);全文 *
弥散张量成像在颅脑损伤昏迷病人中的应用;张晓峰,等;《国际神经病学神经外科学杂志》;20190430;第36卷(第4期);全文 *

Also Published As

Publication number Publication date
CN110801203A (en) 2020-02-18

Similar Documents

Publication Publication Date Title
CN110335297B (en) Point cloud registration method based on feature extraction
Bookstein et al. A feature space for edgels in images with landmarks
CN102782723B (en) Position and direction estimation method and equipment thereof
Gu et al. GPU-based fast gamma index calculation
EP0313586A1 (en) A method for global blending of computer modeled solid objects using a convolution integral
Kenobi et al. Shape curves and geodesic modelling
CN109101741B (en) Complex surface detection self-adaptive sampling method based on triangular mesh simplification
CN108416801B (en) Har-SURF-RAN characteristic point matching method for stereoscopic vision three-dimensional reconstruction
CN109685841B (en) Registration method and system of three-dimensional model and point cloud
Franke et al. Least squares surface approximation to scattered data using multiquadratic functions
Rouhani et al. Implicit polynomial representation through a fast fitting error estimation
CN111859825A (en) Method and equipment for simulating unsteady non-pressure flow numerical value with arbitrary flow-solid interface
CN110801203B (en) Human cranial nerve fiber tracking method based on local features
CN108364268A (en) A kind of single frames bar graph phase recovery method and device
CN105389476B (en) The interpolation algorithm of IMRT intended dose data based on Gradient Features
CN104700368A (en) Self-adaptive sliding method of displacement field of digital image relevant method based on kernel function
CN113658194A (en) Point cloud splicing method and device based on reference object and storage medium
TWI406189B (en) Method for constructing triangular grids of point clouds
CN105387826A (en) method and apparatus for quantifying dimensional variations and process capability
CN115147471A (en) Laser point cloud automatic registration method based on curvature density characteristics
Kleban et al. Finite-size effects at 2D Ising critical points via conformal mapping
US20140044332A1 (en) Transformation method for diffusion spectrum imaging using large deformation diffeomorphic metric mapping
CN114022526A (en) SAC-IA point cloud registration method based on three-dimensional shape context
Wang et al. Brain surface conformal parameterization with the Ricci flow
Hoffmann et al. Robustness of topological defects in discrete domains

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant